Hi, thanks for making your works publicly available. I enjoyed going through your paper and "loss" code. I found a couple of things unclear for me regarding the loss, I will be glad if you will bring some clarifications.
In your paper, you are mentioning that q is a continuous target that has a domain of [0,1]. But in the Eq 7,8,9 in your system equations, you are conditioning your equations based on q> 0 . I have checked your code where you are implementing the loss, as far as I understand you are conditioning your equations based on the condition of if q ∈ [u, u+interval]
So in the Eq 7,8,9, do you actually condition based on if the right interval is found? Or is it my wrong misinterpretations?
You can explain it this way. But the range of the generated q is [0,1], and q> 0 only under the condition of the right interval, you can refer to here.
Hi, thanks for making your works publicly available. I enjoyed going through your paper and "loss" code. I found a couple of things unclear for me regarding the loss, I will be glad if you will bring some clarifications.
In your paper, you are mentioning that q is a continuous target that has a domain of [0,1]. But in the Eq 7,8,9 in your system equations, you are conditioning your equations based on q > 0 . I have checked your code where you are implementing the loss, as far as I understand you are conditioning your equations based on the condition of if q ∈ [u, u+interval]
So in the Eq 7,8,9, do you actually condition based on if the right interval is found? Or is it my wrong misinterpretations?